The Onyx approach to impurity identification and control

The purity of an active pharmaceutical ingredient (API) is one of the many quality attributes that require control to ensure patient safety. As impurities are considered to be any substance in an API that is not the API itself, the term “impurity” can refer to a wide range of chemical species. Due to this broad scope, separate guidelines have been issued by the ICH concerning all classes of impurity in new drug substances:

• Organic and inorganic (Q3A)
• Residual solvents (Q3C)
• Metals (Q3D).

In this blog, Kirsty Milne discusses Onyx’s approach to the identification and control of organic impurities.

If you’re interested in a more detailed discussion on the impurity strategies for inorganic impurities, residual solvents and metals, read our other blog: “Enabling the synthesis of clinical API: Onyx’s approach to defining phase 1 drug substance specifications.” 

Origin of impurities in a chemical process

No reaction is perfect, and almost all chemical processes result in the formation of impurities. These can range from: 

• Unreacted starting materials, reagents and catalysts 
• Reaction intermediates and by-products
  • Activated esters, ureas and HOBt from amide couplings
  • Imines and boric acids from reductive aminations
• Undesired side reactions
  • Hydrolysis, transesterification or dimerization
  • Reaction with impurities found within raw materials 
    • Enantiomers, diastereomers or regioisomers
• Degradation products on storage.

The Onyx approach: Predicting impurities from the outset

Each reaction has its own unique set of potential side reactions and, therefore, impurity profile. The current ICH Q3A guidelines state that impurities which are greater than the identification threshold (the lower of either 0.10% or 1 mg/day for a drug with up to a maximum daily dose of 2 g/day or 0.05% for a drug with a daily dose over 2 g/day) in the final API should be identified (although this is not always possible). 

At Onyx, we aim to identify any impurities that are >0.1% from early development. To achieve this, it is important to fully understand the chemical transformation taking place in a given synthetic step and thereby highlight any potential side reactions or by-products that may occur. This logic can also be applied to the starting materials/reagents employed in the reaction, as these may also be a source of impurities, which may or may not react under the reaction conditions. Having this knowledge in hand at the start of development allows for easier detection of these species, typically by LC-MS and 1H nuclear magnetic resonance (NMR). Once an impurity has been observed, it may be possible to minimise its formation through modification of the reaction conditions, or if this is not possible, an additional purification process may be developed to provide the desired control over a given impurity. By applying this strategy to each stage of a manufacturing process, both the number and level of impurities present in the final API are minimised. 

Fate and purge: mapping impurities across the process

The fate of these impurities can be tracked throughout the whole process by following the reaction with LC-MS. One of our aims in supporting this work is the development of a suitable analytical method which can separate all the starting materials, intermediates, impurities and the API. This allows for the production of process impurity tables, which makes tracking the starting materials, intermediates and impurities across multiple stages easier. This approach allows our chemists to identify where problem impurities arise and at which stage in the process control strategies should be applied to ensure the most efficient purge possible.

For example, if an impurity forms in stage 1 but is unreactive and purged during stage 2, it is more efficient to carry the impurity through and purge at stage 2, rather than add an extra processing step in stage 1.

Some of the key approaches our chemists would utilise are:

• Optimisation of reaction conditions to form less of the impurity
• Development of a suitable work-up procedure 
• Development of a suitable purification
• Changing the synthetic approach to avoid forming the impurity.

Strategies for isolation and synthesis when impurities persist

Nonetheless, some impurities may still be present in the API above the thresholds described above and, therefore, require formal identification. One approach would be to isolate the impurity from either the reaction products or from recrystallisation liquors, through column chromatography or preparative HPLC. It is also possible to modify the reaction conditions to push the reaction into forming more of the impurity to increase the chances of isolation. While this usually doesn’t provide large quantities of material, it is often sufficient to isolate enough of the impurity for characterisation by LC-MS and a variety of 1D (1H , 13C, DEPT 135, etc) and 2D (COSY, HMBC, HSQC, etc) NMR experiments or by single crystal x-ray diffraction and to establish a relative retention time (RRT) with regards to the drug substance. 

Should it not be possible to isolate the impurity by this method, a synthetic route to the proposed structure will be identified and the postulated impurity synthesised and directly compared to the data obtained with either the authentic impurity or by spiking into an HPLC sample of the API and comparing the RRT to that of the known impurity. This approach will usually generate larger quantities of the impurity, which allows for the production of a formal reference standard, where the material is fully characterised by HPLC, NMR, MS and IR spectroscopy. 

Should an authentic sample be unavailable, it is also possible to compare the obtained NMR data with that of a reference standard of a closely related substance (typically the drug substance or an intermediate). From this, it is possible to establish which signals have shifted or are missing, which could help pinpoint where the new bonds have formed in a side reaction. 

Mitigating mutagenic risk: Identifying and controlling PMIs

Once the structure of an impurity has been confirmed, it allows for a number of approaches towards monitoring and control. If the structure of the impurity alerts as a potential mutagenic impurity (PMI),  it is possible to start in silico screening or, if required, send material for Ames testing. This allows for suitable limits to be set according to ICH M7 guidance, as well as suitable control measures and analytical methods to be put in place during processing to ensure those limits are quantifiable and achievable. 

Onyx’s approach to PMI identification, screening and control measures is discussed in more detail in our previous blog: Strategies for Control of Potential Mutagenic Impurities (PMIs) in Pharmaceutical Synthesis.

Quantification and analytical validation

It is essential to confirm that the impurity can be adequately quantified using a given analytical method. The method should be capable of separating all of the known impurities from both the API and each other, as well as having the sensitivity required to accurately quantify those impurities at low levels. This is especially important if the material alerts as a PMI, as the permitted limits may be lower than for non-mutagenic impurities. Additionally, in HPLC analysis, an impurity may have a higher or lower UV absorbance than the API itself and therefore will have been under- or overestimated in previous analysis. By having an authentic sample of the impurity, the relative UV response can be calculated and confirmed that the level of the impurity is within specification. If this is not the case, redevelopment of the chemistry or the inclusion of an additional purification step may be required. 

Proving control measures through spike studies

Furthermore, fate and purge studies can also be expanded on by carrying out spiking experiments, where the impurity of interest is spiked in at higher levels than typically observed during processing, to stress the control measures put in place. This will prove that those measures will effectively control the levels of impurity. This work will also allow our chemists to determine suitable specifications for intermediates by determining the maximum level of the impurity that can be purged over the whole process, and which measures can be applied should the levels be outside of the desired specifications. If it is not possible to control an impurity to the required level, the impurity could be spiked into a batch at a higher level for additional toxicology testing, which would potentially allow for a higher level to be set while maintaining patient safety. 

A robust impurity strategy from the start

By thorough examination of both chemical literature and real-life batch data, it is possible to identify organic impurities and develop and implement suitable control strategies into the manufacturing process from day one. This strategy helps ensure a smooth transition through the phases of development and allows us to work closely with our customers to produce specifications for their drug substance, which are compliant with the ICH guidelines, achievable and reproducible across multiple batches. All of which goes towards the ultimate goal of providing safe and effective treatments for patients.